Free Viewpoint Television

FTV allows the viewer to control the viewpoint and generate new views of a dynamic scene. This means the focus of attention can be controlled by viewers rather than the director or film maker. However this could ruin dramatic suspense caused by not being able to see the whole picture.

Professor Masayuki Tonimoto, who is a researcher in Japan has been promoting the use of Free Viewpoint Television with his technique of Ray Space Representation. Which is a development of Paul Debevec's 3D image Modelling and rendering techniques.

"This method allows it to render virtual views from arbitrary positions without any geometry information. Furthermore the generated image is photo-realistic."

However this is not the only way to create FTV.

How is it created?

1. Firstly cameras are placed around a subject, be it in a studio or a sports stadium.
2. The data from each of the cameras is the collated into a single codec, none as MVV (Multiview Video) which is then compressed so that it can be sent to the viewing device.
3. The users viewing device then uses the codec to access the relevant views so that new views can interpolated.
4. The new footage (which is a mixture of original footage and interpolated images) is then viewed on either a 2D or 3D display.

I believe that this form of viewing television will break into the mainstream, as i don't think it is suitable for viewing or capturing Drama or soaps. However i do see it being used to capture and view live events, such as; Sporting coverage or Music concerts. It could also find it's way into CCTV.

How bullet time was created in the Matrix

John Gaeta used a mixture of techniques to create the final product that is "Bullet time".

• Firstly each of the "bullet-time" sequences were filmed in a Studio in Australia using Green Screens

• The object (usually an actor) was then surrounded by an array of Still and motion cameras. this is the technique developed by Tim Macmillan. Each still camera took a single frame, either at the same moment or in sequential order. This depended on the desired effect. The Motion cameras were used to record the moments leading up to and after the "bullet time" moment, to give fluidity when moving in and out of the effect.

• Image Interpolation was also used to create extra frames so that they could slow down and speed up the action or the speed in which they panned round an object. This was so that they could have more control over how they edited the footage, while maintaining a high level of quality.

• To create the background to the scenes, they used Paul Debevec's technique of 3D image Modelling and Rendering. They did this by taking photos of the locations they wanted to use, For example the Rooftop locations of Sydney, Australia.

The Finished Product....

http://www.youtube.com/watch?v=vXAXyuShGUo&feature=related

Paul Debevec

Debevec was a researcher in Computer Graphics at the University of Southern California. He is Known for his work in high Dynamic range imaging and [more importantly to "bullet-time"] Image-based Modelling and Rendering.

In 1996, his research thesis was in Photogrammetry. This is where the 3D shape of an object could be collected from a group of stills taken from different angles. This lead to him embarking on a task in 1997 where he created a "Bullet-time" style flyby of UC Berkeley's famous Campanile tower.

http://www.youtube.com/watch?v=RPhGEiM_6lM

This was created by firstly taking pictures of the tower and the surrounding area from all sorts of angles. These were then in-putted into Debevec's Image Rendering software. This then created a 3D model which was then rendered and added to real video footage.

Many of the team that created the Campanile film went on to work on the matrix film, including Paul Debevec himself. this technique was used to create the rooftop backgrounds of many of the "bullet-time" scenes.

http://www.youtube.com/watch?v=WhxbYTMNMxo&feature=related

John Gaeta was quoted saying...

"Once I saw Debevec's movie, I knew that was the path"